JP4383237B2 - Self-position detection device, in-furnace inspection method and in-furnace inspection device using the same - Google Patents

Description

  The present invention relates to a self-position detecting device that autonomously detects the self-position of a moving body, an in-furnace inspection method using the same, and an in-furnace inspection apparatus, and particularly, a moving speed such as a moving body that swims in water. The present invention relates to a self-position detecting apparatus suitable for a slow moving body, an in-furnace inspection method using the same, and an in-furnace inspection apparatus.

  More specifically, the present invention relates to a technology for autonomously detecting the self-position of a moving object in a place where position detection by GPS or PHS using radio waves is difficult, and further, autonomous self-position detection. In particular, the present invention provides an error correction method that can be used even when the speed of a moving object is low.

  Furthermore, the present invention relates to an inspection apparatus and an inspection method for a structure in a reactor pressure vessel. At this time, the inspection apparatus is brought close to the structure in the reactor pressure vessel by remote operation, and the structure is inspected. In particular, the present invention relates to an inspection apparatus and an inspection method for performing preventive maintenance such as cleaning, and more particularly, to a technique for efficiently performing a visual inspection based on an image captured by a camera, so-called VT inspection.

  When inspecting nuclear reactor structures, such as shrouds, in order to prevent exposure to radiation, the inspection equipment must be kept in a state where the reactor has been filled with pure water. A mobile body called ROV (remote control vehicle) is submerged in water, and the inspector remotely monitors the camera image while monitoring and cleaning the target part.

  At this time, an in-furnace visual inspection apparatus using a guide rope has been proposed in order to stabilize an image captured by an inspection device (see, for example, Patent Document 1).

  In addition, as an inspection apparatus specialized for a specific location in a reactor pressure vessel, several proposals have been made regarding the inspection technique for the annulus portion of the reactor (see, for example, Patent Document 2 and Patent Document 3).

  Here, an inertial sensor (generally a sensor combining a three-axis accelerometer and a three-axis gyro) is used when self-position detection is required. In this case, a Kalman filter is used as a method for suppressing detection errors. It is common to apply.

  At this time, the position detected by the inertial sensor needs to be calibrated. For this reason, an azimuth detecting device that uses an absolute position detected by GPS has been proposed (see, for example, Patent Document 4). . In this case, a method of detecting the absolute position by PHS can also be applied.

  At this time, an inertial device that detects a stop state and resets the inertial speed has been conventionally known (see, for example, Patent Document 5).

  On the other hand, an underwater vehicle speed measuring device has been proposed in the past in which an error is suppressed by subjecting the propeller rotational speed and the speed detected by the inertial sensor to Kalman filtering (see, for example, Patent Document 6). .

Further, at this time, a posture position detecting device that detects a posture by image processing and detects the posture by mutually correcting the motion of the inertial sensor has been proposed (see, for example, Patent Document 7).
Japanese Patent Laid-Open No. 2003-185783 JP 2003-337192 A JP 2002-286552 A Japanese Patent Application Laid-Open No. 6-288776 JP-A-5-248882 JP-A-7-270176 JP 2000-97637 A

  Among the above-described conventional techniques, the conventional techniques disclosed in Patent Documents 1 to 3 are for the observer to visually grasp the current position and the structure to be inspected, and thus the burden on the observer is large. In addition, once the self-position is lost, it is difficult to search for the self-position.

  Here, in the conventional techniques disclosed in Patent Documents 4 to 7, the self-position is detected using an inertial sensor, and the detection error at this time is reduced using a Kalman filter. There is no need to grasp the position visually.

  However, since the prior art disclosed in Patent Document 4 is premised on absolute position detection by GPS, there is a problem that it cannot be applied in an environment where GPS radio waves cannot be received such as indoors. The conventional technique disclosed in Document 5 uses a vehicle speed sensor for detecting a stop state, and therefore has a problem that it cannot be applied except when a wheel such as a ground traveling body is provided.

  Moreover, in the case of the prior art disclosed by patent document 6, since the moving speed detected with the inertial sensor is correct | amended using the rotational speed of the propelling device (screw propelling device) of a moving body, the inspection in a nuclear reactor is carried out. When the navigation speed is low as in the apparatus, it is difficult to detect the accurate navigation speed from the rotation speed of the propeller, and there is a problem in maintaining accuracy.

  Furthermore, in the prior art disclosed in Patent Document 7, since the motion is detected by image processing and the posture is detected by mutual correction with the motion of the inertial sensor, the absolute position of itself cannot be calculated in the end. Since there is no absolute position error correction means, there is a problem that it cannot be applied as a self-position detecting device.

  The present invention solves the above problems and provides a technique for detecting the self-position with high accuracy in a place where absolute position detection using radio waves such as GPS and PHS is difficult. It is an object of the present invention to provide an inspection apparatus that can be used even when the speed of a moving object to be detected is low.

  Accordingly, an object of the present invention is to provide a self-position detecting device capable of autonomously and accurately detecting the self-position regardless of the environment, a furnace inspection method and a furnace inspection apparatus using the same.

According to a first aspect of the present invention for achieving the above object, in the self-position detecting device of the type for calculating the inertial movement amount of the moving body by the inertial movement amount detecting means and detecting the position of the moving body from the inertial movement amount, A camera mounted on the moving body such that the front of the moving body is a field of view; camera relative movement amount detecting means for calculating a camera relative movement amount of the moving body based on image information captured by the camera; the position of the movable body, represented by the inertial movement amount, and a self-location update means for correcting the position of the movable body which is represented by the camera relative movement is provided, said self-location update means An inertial speed standard deviation calculating means for calculating an inertial speed deviation from a time change of the inertial movement amount, and an image speed standard deviation calculation for calculating an image speed deviation from the time change of the camera relative movement amount. It comprises means, and gain updating means for updating the Kalman gain based on the image velocity difference between the inertial speed deviation, an absolute position updating means for updating the inertial movement amount based on the output of the gain update unit It is characterized by this.

Next, a second invention for achieving the above object is the self-position detection device according to claim 1, wherein the inertial movement amount detection means includes an acceleration detection means for detecting acceleration, and an angular velocity for detecting angular velocity. It comprises a detecting means, an axis speed calculating means for calculating the speed of each axis from the acceleration, and a rotation angle calculating means for calculating a rotation angle from the angular speed .

According to a third aspect of the invention for achieving the above object, in the self-position detecting device according to claim 1, the self -position updating means includes a map storage means in which the three-dimensional structure information of the detection target is stored. The contact determination unit that determines interference based on the self-position of the detection object updated by the absolute position update unit and the three-dimensional structure information read from the map storage unit, and the contact determination unit determines that there is interference. And an absolute position correcting means for correcting the self position .

Furthermore, a fourth invention for achieving the above object is the self-position detection device according to claim 1 , wherein the movable body includes marker laser irradiation means for irradiating a laser parallel to the optical axis of the camera, An image center for calculating a distance from the camera to an object imaged at the center in the image based on a position of the laser light image in the image imaged by the camera A distance calculating means is provided.

Similarly, a fifth invention for achieving the above object is to inspect the inside of the nuclear reactor using the in-core inspection means on which any of the self-position detecting devices according to claims 1 to 4 is mounted. In the in-furnace inspection method , the in- furnace inspection means includes a position / orientation control means and a remote control means for operating the position / orientation control means, and is a moving body capable of arbitrarily navigating underwater. After removing any of the control rod guide tubes in the reactor from the core support plate and pulling them up from the reactor together with the control rods and the fuel assembly, the control rod guide tube was attached to the in-core inspection means. The inspection is performed by inserting into the lower part of the core support plate through the hole of the core support plate .

Similarly, according to a sixth aspect of the present invention for achieving the above object, in the in-reactor inspection method according to claim 5 , the self-position detecting device causes the movable body to contact an upper portion of a CRD housing in the reactor. The position is taken in as a reference position of the moving body .

Similarly, a seventh invention for achieving the above object is to inspect the inside of the reactor using the in-reactor inspecting means on which any of the self-position detecting devices according to claims 1 to 4 is mounted. In the in-furnace inspection apparatus, the in- furnace inspection means includes a position / orientation control means and a remote operation means for operating the position / orientation control means, and is a movable body capable of arbitrarily navigating underwater, the self-position An update means, an inspection path input means for inputting an inspection target path, an inspection target setting means for comparing the self position as an output of the self-position detecting device with the inspection target path, and setting an inspection target for the time being, An inspection apparatus automatic control means for comparing the self-position and the inspection target and controlling the position and orientation control means is provided .

According to the first aspect of the present invention, the relative position detection error due to the inertial sensor can be reduced even in an area where radio waves cannot be used or where an external position detection infrastructure such as a camera or an ultrasonic oscillator cannot be installed. High absolute self-position detection is possible.

Furthermore, according to the first aspect of the present invention, since it is possible to provide means capable of associating image information and inertial sensor information and reducing the relative position detection error of the inertial sensor, highly accurate absolute self-position detection is possible.

According to the second aspect of the present invention, since the speed and position are further calculated from the output of the inertial sensor and associated with the speed detected in the second aspect of the present invention, the same effect as the first aspect can be obtained.

According to the third aspect of the present invention, since the absolute error of the self-position detected by the inertial sensor and the image sensor can be corrected and the accumulated position detection error can be removed, the same effect as the first aspect can be obtained. .

According to the fourth aspect of the present invention, the distance from the single image information only to the object can be calculated, and the same effect as in the first aspect can be obtained.

According to the fifth aspect of the present invention, the self-position can be automatically detected in the in-reactor inspection, and an efficient inspection becomes possible.

According to the sixth aspect of the present invention, since the reference point of the position in the nuclear reactor can be obtained correctly, efficient inspection can be performed.

According to the seventh aspect of the present invention, since the inspection in the nuclear reactor can be automated, the burden on the worker is reduced.

  Hereinafter, a self-position detecting apparatus according to the present invention, an in-furnace inspection method and an in-furnace inspection apparatus using the same will be described in detail with reference to the illustrated embodiments.

  FIG. 1 shows an embodiment of the present invention. In this embodiment, the present invention is applied to an in-core inspection apparatus for visual inspection in a nuclear reactor, and the self-position detection apparatus according to the present invention is mounted on the inspection apparatus. Thus, the burden on the inspection worker can be reduced. Therefore, as shown in FIG. 1, an acceleration detecting means 101, an angular velocity detecting means 102, and a camera (television camera) 107 are provided. ing.

  Here, first, the acceleration detection means 101 and the angular velocity detection means 102 function to detect linear acceleration and angular velocity acting on the inspection apparatus, and the detected linear acceleration and angular velocity are supplied to the inertial sensor signal processing means 103.

  Therefore, the inertial sensor signal processing unit 103 calculates the inertial movement amount IM from the input linear acceleration and angular velocity, and supplies them to the self-position updating unit 104 and the image processing unit 110.

  On the other hand, the camera 107 performs an imaging operation using the front of the inspection apparatus as an imaging field of view, and serves to supply the image data converted into electronic information by the image information acquisition unit 108 to the image processing unit 110.

  At this time, the laser marker irradiating means 109 is used to irradiate the subject with horizontal linear light (marker laser light), which will be described in detail later.

  Therefore, the image processing unit 110 that has received this image data calculates the camera relative movement amount OM using the rotation angle calculated by the inertial sensor signal processing unit 103 and supplies it to the self-position detection unit 104. .

  Then, the self-position updating unit 104 updates the self-position using the inertial movement amount IM and the camera relative movement amount OM, and transmits the result to the absolute position display unit 106 via the signal transmission unit 105. This makes it possible for the inspection worker to grasp the position of the inspection device.

  Here, the details of the inertial sensor signal processing means 103 will be described. Among the means included in the inertial sensor signal processing means 103, first, the axial speed calculation means 111 determines x, y, and x from the acceleration signal supplied from the acceleration detection means 101. The speed (axis speed) in each axial direction of z is calculated.

  Next, the rotation angle calculation unit 112 calculates rotation angles around the x, y, and z axes from the angular velocity signal supplied from the angular velocity detection unit 102. Then, the inertial movement amount calculating means 113 calculates a three-dimensional movement amount, that is, an inertial movement amount IM using the axial speed and the rotation angle.

  Here, the details of the image processing means 110 will be described. Among the means included in the image processing means 110, the image center distance calculating means 114 first uses the image information supplied from the image information acquiring means 108, and from now on. The distance between the subject near the center of the image and the camera is calculated.

  Next, the image change amount calculation means 115 uses the image information and the rotation angle inertial movement amount, and calculates a two-dimensional movement amount of the image from these information. Then, the camera relative movement amount calculation means 116 calculates a three-dimensional movement amount, that is, a camera relative movement amount OM, using the distance between the image and the subject and the two-dimensional movement amount of the image.

  Next, the self-position updating unit 104 will be described. Among the units included therein, the inertia velocity calculation unit 117 first differentiates the inertia movement amount IM calculated by the inertia movement amount calculation unit 113. It works to calculate the inertia speed. The result is recorded in the inertia speed standard deviation calculating means 119, where the standard deviation of the inertia speed from a predetermined time to the present is calculated.

  Here, the predetermined time is a time that is arbitrarily set at a point in time when it is considered that the usage (operation) mode has changed, for example, when the inspection apparatus is moved and during the inspection in the furnace inspection.

  Similarly, the image speed calculation unit 118 serves to calculate the image speed by differentiating the camera relative movement amount OM calculated by the camera relative movement amount calculation unit 116. The result is recorded in the image speed standard deviation calculating means 120, where the standard deviation of the image speed from a predetermined time to the present is calculated.

  Next, the gain updating unit 121 functions to update the Kalman gain based on the standard deviation calculated by the inertia velocity standard deviation calculating unit 119 and the image speed standard deviation calculating unit 120, and the absolute position updating unit 122 The absolute position is updated based on the gain and the previous position stored in the absolute position storing means 126.

  Here, when self-position detection is started, an initialization command is input from the image display means 106 to the initialization means 127, whereby the absolute position and the rotation angle are initialized.

  Next, an underwater electrophoresis type in-furnace inspection apparatus equipped with the self-position detection apparatus shown in FIG. 1 will be described with reference to FIG. Here, FIG. 2 shows an in-furnace inspection apparatus 201, in which an inertial sensor unit 202 is first installed, in which an acceleration sensor 203 and a gyro 204 are built. Here, the installation direction of the acceleration sensor 203 and the gyro 204 at this time will be described later.

  Next, in this in-furnace inspection apparatus 201, an image capturing unit 205 is installed, and a camera 206 (corresponding to 107 in FIG. 1) and a marker laser 207 (corresponding to 109 in FIG. 1) are also installed. These are connected to the image capturing unit 205, whereby the image signal captured by the camera 206 is captured by the image capturing unit 205, and at the same time, the marker laser 207 is driven.

  Further, the in-furnace inspection apparatus 201 is provided with an attitude position control unit 208 and a three-direction thruster 209 connected thereto, so that the attitude and position of the in-furnace inspection apparatus 201 are controlled. It has become.

  The in-furnace inspection apparatus 201 is further provided with a ROV-side signal transmission unit 210, to which an inertial sensor unit 202, an image capturing unit 205, and an attitude position control unit 208 are connected.

  At this time, the signal transmission line 211 is connected to the ROV side signal transmission unit 210, whereby the ROV side signal transmission unit 210 is connected to the control device 213 via the control device side signal transmission unit 212. .

  Therefore, after moving the in-core inspection device 201 into the reactor to be inspected and swimming in water as an underwater electrophoresis ROV, the in-core inspection device 201 transmits a signal from the control device 213 outside the reactor. It will be remotely controlled via line 211.

  Next, the arrangement direction of the acceleration sensor 203 and the gyro 204 built in the inertial sensor unit 202 will be described with reference to FIG.

Now, for each of the three-dimensional axes shown in FIG. 3, the X-axis is the front-rear direction of the in-furnace inspection apparatus 201, the Y-axis is the left-right direction, and the Z-axis is the up-down direction. Then, in this case, as shown in the figure, the roll angle θ _R is around the X axis, the pitch angle θ _P is around the Y axis, and the yaw angle θ _Y is around the Z axis.

  First, the triaxial acceleration sensor 301 is arranged such that the detection axes in the three directions are parallel to the X axis, the Y axis, and the Z axis, respectively. Then, the three-axis acceleration sensor 301 detects three types of axial accelerations of the X axis, the Y axis, and the Z axis.

  Next, the gyros 302, 303, and 304 are arranged so that each detection axis is parallel to the X axis, the Y axis, and the Z axis. Then, the angular velocities around the X-axis, Y-axis, and Z-axis are detected from these gyros 302, 303, and 304, respectively.

  Therefore, the inertial sensor signal processing means 103 in FIG. 1 integrates the acceleration signal in the three-axis direction detected by the acceleration detection means 101 (which is the three-axis acceleration sensor 301 in FIG. 3) with time integration by the axial speed calculation means 111. The velocity in each axis direction is calculated, and the angular velocities around the three axes detected by the angular velocity detection means 102 (which are gyroscopes 302, 303, and 304 in FIG. 3) are time-integrated by the rotation angle calculation means 112 and are Calculate the rotation angle.

  Then, the relative movement amount is calculated by the inertial movement amount calculation means 113 from the speed in each axial direction and the rotation angle around the angular axis.

  Next, the absolute position display means 106 will be described with reference to FIG. 4. First, it is provided with a display device, and the display screen 401 includes an operation unit 402, a horizontal position display unit 403, and a vertical position. Each display area of the display unit 404 and the camera image display unit 405 can be displayed. In this embodiment, a function for recording and reproducing these histories is further provided.

  Here, the operation unit 402 includes monitoring and history reproduction switching, display size switching between the horizontal position display unit 403 and vertical position display unit 404, camera switching, target setting, target display, and other icon displays, and notification. The display for is made.

  Next, the operation of this embodiment will be described with reference to FIG. 5. FIG. 5 is a flowchart showing the entire process. When ROV position detection is started (step 1), first, the initial position is set. And an initial orientation are input (step 2). This is due to the initialization command from the initialization means 127 of FIG. 1, and the absolute position and rotation angle are initialized by the method described later in FIG.

  Next, until an end command is input (step 3), an inertial movement amount calculation process (step 4), a camera relative movement amount calculation process (step 5), an absolute position update process (step 6), and an absolute position output (step 7). ) Repeat each process. When the end command is input, the ROV position detection is ended (step 8).

  Next, the method for giving the initial position in step 2 of FIG. 5 will be described with reference to FIG. 6. This FIG. 6 shows a simplified structure inside the nuclear reactor. The CRD stub 602 and the CRD housing 603 are installed on the control rod guide tube 604, and the control rod guide tube 604 is inserted into the core support plate 605 from a hole (installation hole). Inserted and fixed to the CRD housing 603.

  The in-furnace inspection apparatus 607 (201 in FIG. 2) according to this embodiment pulls out one control rod guide tube 604 in advance, inserts it through the hole 606, and temporarily stops it on the CRD housing. Then, the position stopped at this time is given as the initial position.

  In this embodiment, it is assumed that the operator (the person who performs the inspection work) inputs the three-dimensional coordinates of the initial position every time the inspection is performed. The position may be read and given in accordance with the conversion instruction.

  Next, a method for giving an initial orientation in step 2 of FIG. 5 will be described with reference to FIG. Here, FIG. 7 is a simplified view of the bottom of the reactor as viewed from above, and the control rod guide tube 703 is provided in the control rod guide tube installation hole 702 opened in the core support plate 701 (605 in FIG. 6). Since (604 in FIG. 6) is installed, the in-furnace inspection device 705 (607 in FIG. 6) is inserted through the hole 704 (606 in FIG. 6) from which the control rod guide tube 703 is pulled out.

  At this time, the operator observes from above and monitors the output of the image center distance calculation means 114, and adjusts the attitude direction of the in-core inspection device 707 with the thruster 209 (FIG. 2) so that the value is minimized. . Then, when the output of the image center distance calculating means 114 becomes the minimum, the rotation angle is initialized and the position detection is started.

  The method for giving the initial rotation angle is to monitor the front with a camera attached to the in-furnace inspection device 607 and visually check the nearest control rod guide tube, for example, the control rod guide tube 706 on the display surface. There is also a method of controlling the posture direction so that is in front.

  Next, the inertial movement amount calculation processing in step 4 of FIG. 5 will be described in detail with reference to FIG. Here, when processing is first started in step 1, first, an acceleration signal and an angular velocity signal are input (step 2). Next, the rotation angle is calculated for each of the three axes by the procedure from Step 4 to Step 5 (Step 3).

  First, the offset voltage is subtracted from the angular velocity signal (step 4). This offset voltage is normally shown as a constant value as a gyro-specific spec. In this embodiment, a DC component is extracted from an angular velocity signal using a low-pass filter, and this is used as a reference voltage.

Next, the signal obtained by subtracting the offset voltage from this angular velocity signal is integrated to detect its posture angle (θ _R , θ _P , θ _Y ), and the unit time rotation angle (ω _R , ω _P , ω _Y ) are calculated (step 5). Next, using the posture angles (θ _R , θ _P , θ _Y ), the unit vector (g _x , g _y , g _z ) in the gravity direction is expressed in the self-coordinate system by the equation (1) (step 7). ).

Next, using the acceleration data, the processing from step 8 to step 10 is executed to calculate the inertial movement amount of each axis (step 7). Therefore, first, basic processing of the measured acceleration data (A _x , A _y , A _z ) is executed, and necessary acceleration signals (AC _x , AC _y , AC _z ) are extracted.

At this time, if the gains inherent to the sensors of the respective axes are G _x , G _y , G _z , and the offsets are Vo _x , Vo _y , Vo _z , and the gravitational acceleration is GV, equation (2) is obtained.

  Therefore, the acceleration of each axis calculated by the equation (2) is integrated to calculate the axis speed (step 9). However, if it is determined that the camera relative movement amount (described later in step 9 in FIG. 9) at the previous time is zero, that is, it is determined that the camera is not moving, the axis speed in this step 9 is zero for all axes. .

  By integrating the obtained shaft velocities of the respective axes, an inertial movement amount is calculated (step 10). Finally, the amount of inertial movement and the unit time rotation angle are output (step 11), and the process is terminated (step 12).

  Next, the camera relative movement amount calculation processing in step 5 of FIG. 5 will be described in detail with reference to FIG. When the process is started (step 1), a camera image is input first (step 2). Next, the image center distance is calculated by the processing from step 4 to step 7 (step 3). A method for calculating the image center distance will be described later with reference to FIG.

  Next, as the image change amount calculation processing, the two-dimensional shift amount of the image is calculated by the processing from step 8 to step 20 (step 7). A method for calculating the image change amount will be described later with reference to FIG.

  Next, using the image center distance and the image change amount, a relative movement amount from the image information is calculated (step 21), the result is output (step 22), and the process ends (step 23).

  Here, the image center distance calculation processing in step 3 of FIG. 9 will be described in detail with reference to FIG. This process calculates the distance between the camera 1001 (corresponding to 206 in FIG. 2) and the object in the vicinity of the center of the imaging field. For this reason, first, the camera 1001 and the marker laser light source 1002 (FIG. 2) are calculated. (Corresponding to 207) at a distance d in parallel, and the position of the marker laser beam 1004 irradiated on the measuring object 1003 is analyzed to determine the distance L.

That is, in FIG. 10, in the obtained camera image 1005, the number of pixels 1008 between the image 1006 by the marker laser beam and the center line 1007 in the Y-axis direction of the camera image is pic1, and the number of pixels in the Y-axis direction of the image 1005 The distance L from the camera 1001 to the object 1003 can be calculated by the equation (3) where the half value 1009 of the camera is pic2 and the viewing angle in the Y direction of the camera 1001 is α.

  Therefore, in step 3, the image in which the laser line (image 1006 by the marker laser beam) is reflected is binarized (step 4), and the distance between the image center and the laser line is detected (step 5). Then, the distance L between the object and the camera is calculated using equation (3).

  Next, as an image change amount calculation process, a method for calculating the amount of movement between the image and the parallel plane will be described in detail with reference to FIG. Now, assuming that the current time is i and reading the window image 1101 at the immediately preceding time i-1, the object 1102 is shown in the window image 1101.

  Next, the central portion 1104 is cut out from all the current images 1103 to obtain a window image 1105 at the current time. When the two-dimensional cross-correlation between the window image at time i and the immediately preceding time i−1 is obtained, a shift amount (ξ, η) corresponding to the difference between the object 1102 and the object 1106 that are the same object is calculated. be able to.

  The calculation procedure at this time will be described according to Step 8 to Step 20 in FIG. First, the previous window image Ii-1 and the current image are input (steps 8 and 9). Since the current image here includes not only the change in the self position but also the influence of the change in the posture of the self, the influence is corrected using the posture angle detected by the inertial sensor (step 10). .

First, posture angles (θ _R , θ _P , θ _Y ) are input (step 11), and the difference between the posture angles and the immediately preceding rotation angle is calculated as a rotation angle change amount (ω _R , ω _P , ω _Y ) (step 12). ). Here, in the embodiment of FIG. 2, since the central axis of the camera 206 coincides with the X axis, the rotation angle change amount of the image is the rotation angle change amount around the X axis, that is, ω _R.

Therefore, if the coordinates of an arbitrary point of the image are (X, Y), the coordinates (X ′, Y ′) after rotation are calculated by the following equation (4).

Next, the entire image is rotated using the equation (4) (step 13). Next, the effects of rotation around the Y axis and around the Z axis are corrected. At this time, the rotation around the Y axis and around the Z axis (ω _P , ω _Y ) is vertical and horizontal on the image, respectively. It can be approximated as a parallel movement.

That is, if the coordinates before translation are (X ′, Y ′) and the coordinates after translation are (X ″, Y ″), these are the distances L to the object calculated by equation (3). Therefore, it can be approximated by equation (5).

  Therefore, the entire image is shifted using the equation (5) to complete the correction of the posture angle (step 14). Next, the process of shifting the image two-dimensionally within the set ranges (ξ1 → ξ2, η1 → η2) and calculating the two-dimensional cross-correlation coefficient is repeated (step 15).

  Here, ξ and η are the movement numbers of the respective pixels in the X direction and the Y direction. Therefore, first, the image is shifted by (ξ, η) (step 16), and the window Ii in the central portion is cut out (step 17).

Then, the average intensities <Ii-1> and <Ii> are calculated for the immediately preceding window Ii-1 and the current Ii image (step 18), and these are used to calculate the cross-correlation coefficient according to equation (6). C (ξ, η) is calculated (step 19). In this case, (j, k) in the equation (6) indicates coordinates in the window.

  Thus, using equation (6), all cross-correlation coefficients C (ξ, η) in the set ranges (ξ1 → ξ2, η1 → η2) are obtained, and C (ξ, η) is minimized ( ξ, η) is calculated as an image change amount, that is, an amount of translation of the center of the image (step 20).

Next, the amount of translation in the plane perpendicular to the optical axis of the camera is calculated from the amount of translation of the image thus calculated. In this case, using the half-value pic2 of the number of pixels in the vertical direction of the image shown in FIG. 10, the viewing angle α of the camera, and the distance L between the camera and the image center object calculated in step 6, Then, the amount of parallel movement (ΔXP, ΔYP) of the camera is calculated (step 21).

Further, the amount of change in the distance between the camera and the object at the center of the image, that is, the amount of movement ΔL in the optical axis direction of the camera is calculated from the difference in the distance L obtained in step 6, and the three-dimensional self-position change amount Is calculated as a camera relative movement amount ΔP from equation (8) based on the mean square with the amount of change obtained in step 21 (step 22).

  Then, the result is output (step 23), and the process is terminated (step 24).

  Next, the flow of the absolute position update process in step 6 of FIG. 5 will be described in detail with reference to FIG. When the process is started (step 1), the inertia speed standard deviation is calculated from step 3 to step 5 (step 2).

  First, the amount of inertial movement is taken (step 3), and the difference from the previous value is taken to obtain the inertial velocity (step 4). At this time, this value is accumulated from the initial state, and the standard deviation is sequentially updated (step 5). Then, by the same process, the image speed standard deviation is calculated from step 7 to step 9 (step 6).

Next, the Kalman gain is updated by the following method (step 10). First, the equation of motion represented by equation (9) and the observation equation represented by equation (10) are constructed.

Here, Xk represents a three-dimensional self-position, uk is an input from the inertial sensor, and as shown in the equation (11), the inertial movement amount ( _Px , _Py , _Pz ) and the unit time rotation It is a vector whose elements are angles (ω _R , ω _P , ω _Y ).

  Here, Yk represents the three-dimensional position obtained from the image information, and A, B, and C are observation matrices for associating these pieces of information, and are determined in advance using the inspection apparatus of the present invention.

Gk and Hk shown here are Kalman gains, which are calculated by the following equation (12). Note that U and W in the equation (12) are the inertia velocity standard deviation and the image velocity standard deviation calculated in Step 5 and Step 9, respectively.

Therefore, next, the Kalman gains Gk and Hk calculated (updated) and the immediately preceding absolute positions Xk−1 and uk are used to calculate and update the absolute position according to the equation (13) (step 11).

  Thereafter, contact is determined based on the updated self-position, and if it interferes with the structure, the absolute position is corrected (step 12). That is, as will be described below, the processing from step 14 to step 23 is repeated until it is determined that contact is not made at step 13.

  First, contact determination is performed at N points around the absolute position updated in step 11 (step 14). The N point is arbitrarily set in advance in the following equation (14) as a relative position (δx, δy, δz) with respect to the center position.

At this time, (X0, Y0, Z0) are the center coordinates of the mounted device, and are the coordinates updated in step 11, and the surrounding point (Xi, Yi, Zi) is expressed by the following equation (14) Defined in

  Next, the three-dimensional bitmap data of the structure stored in advance is called up. In this process, the presence or absence of a structure is confirmed at the interference check point calculated in step 15, and a return value of 1 or 0 is obtained in step 17. If the return value is 0, it is determined to be non-contact and if it is 1, it is determined to be contact.

  This determination is performed at all N surrounding points, and when it is determined that contact is not made at any point, the contact determination / absolute position correction processing is completed. On the other hand, if contact is determined at any point, after the processing from step 21 to step 23, the processing from step 14 to step 19 is performed again.

First, the absolute position is returned to the position before updating, that is, the position before updating in step 11 (step 21). Then, the rotation angle θ _Y about the Z axis is corrected by a fixed angle δθ _arr in the movement direction of the mounted device (step 22), and the absolute position is updated again by the same process as step 11 (step 23). ).

  After the contact determination / absolute position correction process is performed in this way, the absolute value is stored thereafter (step 24), and the update process is terminated (step 25).

  Note that the correction method from step 21 to step 23 can be applied by other methods as long as the correction amount is within a minimum range and is moved to a position where it does not interfere with the structure.

  For example, hold the amount of inertia movement, rotate the azimuth to the minimum in a range where interference does not occur, store the correction history of the corrected rotation angle, calculate the system error of the rotation angle from the average value of angle correction, and rotate There is also a method of correcting the corner itself, and this method may be applied.

  Therefore, according to one embodiment described above, a Kalman filter is constructed using an inertial sensor and image information to realize self-position detection with a small error, and more effective in-core inspection by an absolute position correction unit. Self-position detection is possible.

  Next, an in-furnace inspection apparatus according to another embodiment of the present invention will be described with reference to FIGS.

  Here, in the embodiment described with reference to FIG. 1, all the signal processing units are mounted on the inspection apparatus and only the final self-position detection result is transmitted. However, in the other embodiments described in FIG. Only the inertial sensor and the image acquisition information are transmitted, and the self-position update and the remote operation are performed on the control device side.

  Here, in this other embodiment, since the means for detecting the absolute azimuth is applied as the method for detecting the angle, the inertial sensor signal processing means is different. Therefore, in the following description of the other embodiment, an explanation will be given with a focus on differences from the already described embodiment.

  First, in FIG. 13, in the case of this other embodiment, acceleration detecting means 1301 (corresponding to 101 in FIG. 2), azimuth detecting means 1302, camera 1303 (same as 107), image information obtaining means 1304 (same as 108). In addition, the position / orientation control means 1305 is mounted on the ROV side, that is, in the in-furnace inspection apparatus 201 shown in FIG.

  The embodiment of FIG. 13 also includes means corresponding to the laser marker irradiation means 109 shown in the embodiment of FIG. 1, but is omitted in FIG.

  The position / orientation control unit 1305 at this time corresponds to the position / orientation control unit 208 and the thruster 208 in FIG. 2, and the ROV-side acceleration detection unit 1301, the direction detection unit 1302, and the image information acquisition unit 1304 include these. Each is connected to a control device 1308 (corresponding to 213 in FIG. 2) via a signal transmission means 1306.

  Therefore, the signal transmission means 1306 here corresponds to a combination of the ROV side signal transmission unit 210, the signal transmission line 211, and the control device side signal transmission unit 212 in FIG.

  As a result, the control device 1307 takes in the information from the acceleration detection means 1301, the orientation detection means 1302, and the image information acquisition means 1304 via the signal transmission means 1306, and the position / attitude control means 1305 receives the signal transmission means. A control signal is transmitted via 1306.

  First, the control device 1307 includes inertial sensor signal processing means 1308. Thus, the acceleration signal and the azimuth signal are processed. Here, the inertial sensor signal processing means 1308 will be described in detail later with reference to FIG.

  The control device 1308 includes a camera relative movement amount calculation unit 1309 and a self position update unit 1310, which are respectively the camera relative movement amount calculation unit 116 and the self position update unit 104 in the embodiment of FIG. It corresponds to.

  First, the camera relative movement amount detection means 1309 calculates a three-dimensional movement amount from the image information. Then, the self-position updating unit 1310 updates the self-position based on information from the image inertial sensor signal processing unit 1308 and the camera relative movement amount detection unit 1309.

  The signal processing procedure by the camera relative movement amount detection unit 1309 and the self position update unit 1310 at this time is the same as the processing procedure by the camera relative movement amount calculation unit 116 and the self position update unit 104 in the embodiment of FIG. Since there is, explanation is omitted.

  Here, as a feature of the embodiment of FIG. 13, the control device 1207 may include an inspection route input unit 1312. Then, prior to the inspection, the operator makes a plan for the route for inspecting the furnace in advance, and inputs the inspection route planned in advance by the inspection route input means 1312. There exists a point which updates the position of the target object used as an inspection target.

  Then, the target position input by the inspection route input means 1312 is sequentially sent to the inspection target setting means 1313 and compared with the self position updated by the self position updating means 1310 so that the traveling direction is determined. It has become.

  Therefore, the orientation (traveling orientation) determined at this time is supplied to the inspection apparatus automatic control means 1305, and according to this, the thruster to be driven is selected from the plurality of thrusters provided in the position and orientation control means 1305. Then, the driving force of the selected thruster is determined and transmitted to the remote control means 1311. Here, the terminal voltage to be applied to the motor of the thruster is determined, and the position of the in-furnace inspection apparatus is controlled.

  Here, in the case of the control device 1307 in the embodiment of FIG. 13, the manual operation input means 1315 is provided. The manual operation control input supplied from now on is supplied to the remote operation means 1311 so that it can cope with an emergency or the like.

  The information of the self-position updating unit 1310, the inspection route input unit 1312, and the inspection target setting unit 1313 is transmitted to the display unit 1316 and displayed.

  Next, the flow of processing in the embodiment of FIG. 13 will be described in detail with reference to FIG. In FIG. 14, when the ROV control is started (step 1), first, the inspection route / final target position is input (step 2). Then, in response to this, the processing from step 4 to step 16 is repeated until the final target position is reached (step 3).

  Here, when the target position is not reached, first, the processing shown in Step 5 to Step 7 is executed, thereby setting the target route (Step 4). First, the target position that should be headed first is set as a provisional target position (step 5), then the three-dimensional bitmap is called (step 6), and then the shortest path that does not interfere with the structure is set. (Step 7).

  As a method for determining the shortest path at this time, a method may be adopted in which the current position and the target position are linearly approximated on a three-dimensional bitmap and corrected from the current position side so as to avoid the structure. However, instead of this, it can be determined as a so-called minimum cost problem using, for example, the Dijkstra method or the Warshall Floyd method.

  Then, the ROV (in this case, the in-furnace inspection apparatus) is driven until the provisional target is reached according to the set shortest distance (step 8). That is, first, as a ROV drive operation (step 9), the drive amount of each drive device (thruster) is set (step 10), and a motor such as a thruster is actually driven (step 11). In this embodiment, since the ROV is an underwater electrophoresis type, the driving device is a thruster and is controlled by this thruster. However, in the case of the self-propelled type, the driving wheel motor is controlled.

  Next, the self-position update means 1310 updates the self-position as described in the self-position update means 104 in the embodiment of FIG. 1 (step 12), and then the current position at the current position with respect to the target position. It is determined whether or not the drive direction is suitable (step 13). If the determination is necessary, the current drive amount of the drive device is adjusted (step 14).

  At this time, as described above, in the embodiment of FIG. 13, a manual stop is also accepted. Therefore, it is next determined whether or not there is a manual stop signal (step 15). If there is a manual stop, the processing routine is ended here even if the final target has not been reached (step 16).

  If it is determined that the final target has been reached by the above control, the ROV control is terminated here (step 17).

  Next, processing by the inertial sensor signal processing means 1308 will be described with reference to FIG. Here, in the embodiment of FIG. 13, acceleration detecting means (acceleration sensor) 1301 and azimuth detecting means (azimuth sensor) 1302 are used to detect the attitude of the ROV. At this time, the azimuth detecting means 1302 includes A magnetic orientation sensor (geomagnetic sensor) is used.

  When the processing of FIG. 15 is started (step 1), first, the signal of the azimuth detecting means (geomagnetic sensor) 1302 is processed to calculate the geomagnetic azimuth (rotation angle from the X axis to magnetic north) (step 2). Therefore, first, a geomagnetic sensor signal is input (step 3).

At this time, the geomagnetic sensor outputs two signals of the X axis and the Y axis, and the azimuth can be calculated by the arc tangent thereof. Therefore, basic processing is performed by subtracting the offset from the geomagnetic signal of each axis and multiplying the gain (step 4). Then, if the X-axis and Y-axis signals after the basic processing are Mx and My, respectively, the geomagnetic azimuth angle θ _MAG can be calculated by the following equation (15).

Here, the geomagnetic azimuth angle θ _MAG represents an angle from the X axis to the magnetic north, and indicates a rotation angle of the X axis when a plane including the X axis and the Y axis is projected onto the horizontal plane.

  Next, a signal of the acceleration detecting means 1302 is input (step 6), and thereafter a gravity direction unit vector is calculated (step 7). That is, first, a signal is processed by a low-pass filter having a frequency of 1 Hz or less, an effective value is calculated, and a DC component of acceleration of each axis is extracted (step 8). Note that this step may use the value of the DC component obtained as a result of the Fourier transform, in addition to the processing by the low-pass filter.

  Next, the gravity component is extracted by subtracting the offset of each axis and multiplying the gain. The vector (GxR, GyR, GzR) obtained there is converted into a unit vector to be a gravity direction unit vector (Gx, Gy, Gz) (step 9).

  Next, the gravity direction unit vector (Gx, Gy, Gz) obtained in step 9 is used as a normal vector, and the magnetic north obtained in step 5 is obtained together with the original north with respect to the plane passing through the self position. Projection is performed on the XY axes (step 11).

Next, the Z axis is rotated, and the XY axis of the gravity direction unit vector and the XY axis by projection are combined to obtain a rotation angle θ _Y around the Z axis. Next, the plane is rotated around the Y axis so that the Z axis is vertical, and the rotation angle θ _P is obtained. Finally, the plane is rotated around the X axis so that the plane is horizontal, and the rotation angle θ _R is set (step 12).

Next, a unit time rotation angle (ω _R , ω _P , ω _Y ) is calculated from the rotation angles (θ _R , θ _P , θ _Y ) thus obtained as a difference from the previous update. The processing after the portion for processing the signal of the acceleration detecting means 1301 and calculating the inertial movement amount (step 14 to step 19) is the same as that in the first embodiment already described in FIG. .

  Therefore, according to the embodiment described with reference to FIG. 13, the inspection can be automated and the inspection efficiency can be further improved.

  The present invention relates to a reactor pressure vessel internal structure inspection apparatus, inspection method, and the like, and in particular, in a nuclear reactor pressure vessel internal structure, an inspection device equipped with a camera is brought close by remote operation, and the structure is visually inspected, It is widely applicable to self-position detection of three-dimensional moving objects, such as when performing so-called VT inspection efficiently and performing preventive maintenance such as cleaning, especially in environments where radio waves such as GPS and PHS cannot be used. Are suitable for use on low-speed moving objects.

It is a block block diagram which shows one Embodiment of the self-position detection apparatus which concerns on this invention. It is a block block diagram which shows one Embodiment at the time of applying the self-position detection apparatus which concerns on this invention for the inspection in a furnace. It is explanatory drawing which shows an example of the arrangement | positioning condition of the inertial sensor in one Embodiment of this invention. It is explanatory drawing which shows an example of the display screen in one Embodiment of this invention. It is a flowchart which shows the whole process by one Embodiment of this invention. It is explanatory drawing of the initial position setting method in one Embodiment of this invention. It is explanatory drawing of the initial position setting method in one Embodiment of this invention. 5 is a flowchart detailing an inertial movement amount calculation process in an embodiment of the present invention. 5 is a flowchart detailing a camera relative movement amount calculation process in an embodiment of the present invention. It is explanatory drawing of the objective distance calculation method in one Embodiment of this invention. It is explanatory drawing of the camera relative movement amount calculation method in one Embodiment of this invention. It is the flowchart which detailed the absolute position update process in one Embodiment of this invention. It is a block block diagram which shows other one Embodiment of the self-position detection apparatus which concerns on this invention. It is a flowchart which shows the whole process in other one Embodiment of this invention. It is the flowchart which detailed the inertial movement amount calculation process in other one Embodiment of this invention.

Explanation of symbols

103: Inertial sensor signal processing means 104: Self-position updating means 110: Camera relative movement amount calculating means 119: Inertial speed standard deviation calculating means 120: Image speed standard deviation calculating means

Claims (7)

In the self-position detecting device of the type that calculates the inertial movement amount of the moving body by the inertial movement amount detection means and detects the position of the moving body from the inertial movement amount,
A camera mounted on the moving body such that the front of the moving body is a field of view; camera relative movement amount detecting means for calculating a camera relative movement amount of the moving body based on image information captured by the camera; Self-position updating means for correcting the position of the moving body represented by the inertial movement amount by the position of the moving body represented by the camera relative movement amount ;
The self-position update means
An inertial speed standard deviation calculating means for calculating an inertial speed deviation from a time change of the inertial movement amount;
Image speed standard deviation calculating means for calculating an image speed deviation from a time change of the camera relative movement amount;
A gain updating means for updating a Kalman gain based on the inertia velocity deviation and the image velocity deviation;
A self-position detecting device comprising: absolute position updating means for updating the amount of inertia movement based on the output of the gain updating means . The self-position detecting device according to claim 1,
The inertial movement detection means is
Acceleration detecting means for detecting acceleration;
Angular velocity detection means for detecting angular velocity;
An axis velocity calculating means for calculating the velocity of each axis from the acceleration;
A self-position detecting device, comprising: a rotation angle calculating means for calculating a rotation angle from the angular velocity . The self-position detecting device according to claim 1,
The self-position update means
Map storage means for storing the three-dimensional structure information of the detection object;
Contact determination means for determining interference based on the self-position of the detection object updated by the absolute position update means and the three-dimensional structure information read from the map storage means;
A self-position detecting device comprising: absolute position correcting means for correcting the self-position when the contact determining means determines that there is interference . The self-position detecting device according to claim 1 ,
The moving body includes marker laser irradiation means for irradiating a laser parallel to the optical axis of the camera,
The camera relative movement detecting means is
Image center distance calculating means for calculating a distance from the camera to an object imaged at the center in the image based on the position of the laser light image in the image imaged by the camera . Self-position detecting device characterized by. In the in-core inspection method in which the inside of the nuclear reactor is inspected using the in-core inspecting means in which any of the self-position detecting devices according to claim 1 is mounted .
The in- furnace inspection means includes a position and orientation control means and a remote operation means for operating the position and orientation control means, and is a moving body capable of arbitrarily navigating underwater,
After removing one of the control rod guide tubes in the reactor from the core support plate and pulling up the control rod and the fuel assembly together with the control rod, the control rod guide tube is attached to the in-core inspection means. An in-core inspection method, wherein an inspection is performed by inserting the core support plate into a lower portion of the core support plate through a hole in the core support plate . In the in-furnace inspection method according to claim 5 ,
The in - reactor inspection method, wherein the self-position detecting device takes in the position as a reference position of the moving body when the moving body is brought into contact with an upper portion of a CRD housing in the nuclear reactor . In furnace inspection apparatus designed to inspect nuclear reactor using either the furnace inspection means mounted self-position detecting apparatus according to claim 1 to claim 4, wherein,
The in- furnace inspection means includes a position and orientation control means and a remote operation means for operating the position and orientation control means, and is a moving body capable of arbitrarily navigating underwater,
The self-position updating means compares the self-position, which is the output of the self-position detecting device, with the inspection target path, and sets an inspection target for the time being. And an inspection apparatus automatic control means for controlling the position and orientation control means by comparing the self position with an inspection target .

Images